Ellipse, Hyperbola and Their Conjunction: Arkadiusz Kobiera
Ellipse, Hyperbola and Their Conjunction: Arkadiusz Kobiera
Ellipse, Hyperbola and Their Conjunction: Arkadiusz Kobiera
Arkadiusz Kobiera
arXiv:1805.02111v2 [math.HO] 26 Jan 2019
Abstract
the given curve is an ellipse, then the locus of vertices of the cones is a
hyperbola. The hyperbola has foci which coincidence with the ellipse
vertex of the cones is the ellipse. In the second case, the foci of the
which originate from the cones used for generation of these curves.
1
Introduction
The conical curves are mathematical entities which have been known for
thousands years since the first Menaechmus’ research around 250 B.C. [2].
Anybody who has attempted undergraduate course of geometry knows that
ellipse, hyperbola and parabola are obtained by section of a cone by a plane.
Every book dealing with the this subject has a sketch where the cone is sec-
tioned by planes at various angles, which produces different kinds of conics.
Usually authors start with the cone to produce the conic curve by section.
Then, they use it to prove some facts about the conics. Many books focus
on the curves themselves and their features. Even books which describe the
conics theory in a quite comprehensive way [2, 4, 1] abandon the cone after
the first couple paragraphs or go to quite complex analysis of quadratics. We
may find hundreds of theorems about the curves but the relation between
the cone and the conics is left to the exercise section at best [4] or authors
quickly go to more complex systems of conics in three-dimensional space [2].
Probably the cone seems to be too simple to spent time on this topic, how-
ever we will show that the cone (strictly speaking family of cones) may have
interesting properties as well. Apart of pure geometry, celestial mechanics is
the second field where conics are important – the orbits are conic curves. Un-
fortunately, the books about celestial mechanics say only a few words about
the the cone if any at al. [3, 5]. In this short paper we would like to focus
on the cone and its relation to conic curves which is surprisingly omitted in
books, but interesting.
2
Ellipse and the Cones
Our task is to find the vertex E of the cone S. Apart from the foci,
the ellipse has also two characteristics points: the vertices A and B. The
distances from the vertices to one of the focus e.g. F1 will be noted as
ra = |F1 A|, and rb = |F1 B|. The semi-axes of this ellipse are a = |AJ| and
b = |H1 J| where J is the center of the ellipse. The distance between foci is
3
c = |F1 F2 |. The radii ra and rb define the eccentricity:
ra − rb c
e= = . (1)
ra + rb a
Obviously, we may use any set of these parameters to define the ellipse E,
however we will prefer the radii and focus F1 .
The first question is about the cone: ”Is the cone S unique?” The answer
is in the following lemma:
Proof :
The proof will be explained in a rather quite informal manner. To solve
this exercise let’s reduce the three-dimensional problem to a two-dimensional
problem by considering plane τ which is defined by cone’s axis and foci (or
vertices) of the ellipse. It is shown in Figure 2. We put line a on plane τ .
The line coincidences with the ellipse vertices A and B and the foci F1 , F2
as well. The line a is also an intersection of planes τ and ρ. Note that the
focus points (e.g. F1 ) are points of tangency of a sphere of center O with
the plane ρ. This sphere is called Dandelin’s sphere and it is simultaneously
tangent to the cone. The tangency points of the sphere and the cone create
a circle which defines plane ω [2]. The intersection of planes ω and ρ is line
f. We create also an additional line b on plane ω which is perpendicular to
f and goes through the axis of the cone. The intersection of the Dandelin’s
sphere by the plane τ is a circle with center O. The circle is tangent to lines
4
t1 and t2 which are two elements of the cone. These lines are obtained by
cutting the cone by plane τ . They meet line a at points A and B.
5
when the point E goes to infinity. In this case the ellipse E is obtained as
a section of the cylinder. One may show that the limiting radius is equal to
the minor semi-axes of the ellipse
√
rmax = a 1 − e2 . (2)
If the radius r can be of any length between 0 and rmax then the location
of point E is not unique and its position depends on radius r. Hence, one
can construct an infinite number of cones which may be used to generate the
ellipse E.
If the cone S is not unique, the next question is: ”What is the locus of
the cone vertices E?” First, we calculate the distance from the cone vertex
E to the ellipse vertex B
The second equality results from the fact that BD and BF1 are tangent to
circle O and they have common endpoint B. Obviously, the angles ∠F1 BO
and ∠DBO are equal and right triangles OF1 B and ODB are congruent.
Then segments F1 B and BD are of the same length rb . One can write
similar equations for segment EA
Here we use the fact that triangles HOE and DOE are congruent and tri-
6
angles HOA and F1 OA are congruent as well. Comparison of the above
equation leads to following proposition:
Prof :
Let us calculate the difference of length of two segments EB and EA
7
Figure 3: Hyperbola H.
right branch is for the case where the sphere is tangent at point F2 . If the
radius r of the sphere vanishes to 0, the point E goes toward foci F1 or F2 .
If the sphere’s radius r goes to the maximum value rmax the point E goes to
infinity on the hyperbola’s branches. Asymptotic lines s1 , s2 are the axes of
cylinders which are limiting cases of the cones with vertex in infinity.
Now we can ask reversed question: What is the locus of vertices G of cones
Z which generate the given hyperbole H. One can consider the hyperbola
H which was found in the previous part. This will not reduce generality of
8
our reasoning. We will keep same plane τ where four points are defined A,
B, F1 and F2 . They also define the hyperbola H on the plane τ . Figure 4
shows the situation where the hyperbola is created by sectioning the cone Z
by plane τ . We state the following lemma by analogy to the case of ellipse:
Proof:
The proof is analogous to proof of Lemma 1. First, we reduce the problem
to planimetry by considering the plane ρ (see Figure 5).
9
Figure 5: Section of the cone Z by plane ρ.
10
The next step is finding the locus of vertices G. By analogy to the Propo-
sition 1 we write following proposition:
Proof:
We will look for the relationship between the distances from the vertex
G to the points F1 and F2 . First, we will consider the right triangle O3 P O4
(Figure 5). Point P is the normal projection of point O4 onto segment AO3 .
By using Pythagoras theorem we have
where r2 is radius of the circle O3 . The triangles GKO3 and GM O4 are also
right triangles because the points K and N are points of tangency of the
lines t3 and t4 to the circles O3 and O4 . Hence, one can write
Recalling that
|O3 O4 | = |O3 G| + |O4 G| (10)
11
and substituting equations (8), (9) and (10) to equation (7) the following
equation is obtained
When both sides of equation (11) are divided by |O4 M ||O3 K| we get
The triangles O3 KG and O4 M G are similar because they are right trian-
gles and and angles ∠KGO3 and ∠M GO4 are equal. The second statement
is true because the triangles O4 M G and O4 N G are congruent and angles
∠KGO3 and ∠N GO4 are congurent as well (points G, O3 and O4 lie on
the axis of the cone, hence the segments O3 G and O4 G are co-linear). Let
the measure of angles ∠KO3 G and ∠M O4 G be χ. Simple trigonometrical
relations based on Figure 5 yield:
|KG| |GM |
= tan χ = , (13)
|O3 K| |O4 M |
|O3 G| 1 |O4 G|
= = . (14)
|O3 K| cos χ |O4 M |
The second term of left hand side of equation (12) can be simplified by use
of the two relationship stated above
|KG| |GM | |O3 G| |O4 G| 1
− = (tan χ)2 − = −1. (15)
|O3 K| |O4 M | |O3 K| |O4 M | (cos χ)2
12
Finally, we get the simple equation:
The next step is finding the product r1 r2 . Let’s note that the angle ∠AF1 O3
is equal to ∠AF1 O3 and it is π/2 − ψ. We may say the same about ∠KF1 O3 .
This fact leads to conclusion that the angle ∠N F1 B is equal to
Obviously, the line F1 O4 is the bisector of this angle. Hence, the angle
∠O4 F1 B is equal to ψ. Triangles F1 AO3 and F1 BO4 are similar and we may
write the following proportion:
|O3 A| |F1 B|
= . (18)
|F1 A| |O4 B|
Hence, r1 r2 = ra rb .
We successfully arrived to the conclusion that the sum of length of seg-
ments GK and GM is constant
13
|AB| is equal to ra + rb then
The fact that ra = |AF1 | = |F1 K| = |F2 M | = |BF2 | and equation (21)
allows us to calculate the sum of the distances between vertex G and the foci
F1 and F2
Hence, the sum of distances of the vertex G from foci F1 and F2 is constant
(|AB| = ra + rb )
|GF1 | + |GF2 | = |AB|. (23)
This equation is the simplest form of definition of the ellipse and we proved
the proposition.
Conclusions
14
such simple mathematics may lead to such interesting results and express
the beauty of geometry that is imperfectly shown in Figure 6.
References
[2] Julian L. Coolidge. A History of The Conic Sections and Quadric Sur-
faces. Dover Publications, New York, 1968.
15
[3] Gerhard Beutler. Methods of celestial mechanics. Vol. I, Physical, Mathe-
matical, and Numerical Principles. Astronomy and Astrophysics Library.
Springer, Berlin ; Heidelberg, 2005.
16